1
$\begingroup$

I have the following equation: \begin{equation} f(x,y)\equiv -1296 x^4 \left(4 y^2+9\right)+9 x^2 \left(-128 y^4+864 y^2+2187\right)-y^2 \left(8 y^2+81\right)^2=0 \end{equation}

Which represents a lemniscate (see picture).

I would like to know if this a well-known lemniscate (for example Lemniscate of Gerono, Lemniscate of Bernoulli etc). I have tried to compare with the ones listed on wikipedia, but it seems that none fit this specific equation.

Is $f(x,y)$ a well-known lemniscate? or is there at least a a 'smarter' representation of $f(x,y)$?

edit: in polar coordinate the equation becomes: \begin{equation} -\frac{1}{4} r^6 (3 \sin (\phi )+2 \sin (3 \phi ))^2-\frac{81}{16} r^4 (2 \cos (2 \phi )+1)^2+\frac{6561}{256} r^2 (2 \cos (2 \phi )+1) \end{equation}

enter image description here

$\endgroup$
4
  • $\begingroup$ Can you expand and convert to polar coordinates ? $\endgroup$ – Yves Daoust Nov 16 '20 at 16:59
  • $\begingroup$ The equation looks much too complex to be anything classical. $\endgroup$ – Yves Daoust Nov 17 '20 at 13:37
  • $\begingroup$ The Lemniscates of Bernoulli, Booth, and Gerono are all degree-four algebraic curves, so your degree-six object is definitely not one of them. If you're wondering whether your curve has interesting geometric properties, perhaps you could say something about how it was derived. $\endgroup$ – Blue Nov 17 '20 at 13:45
  • $\begingroup$ Anyway, after dropping a factor $r^2$ you can factor as two equations in $r^2$. Probably one of them has no real roots, otherwise the curve would look more complicated. This should be investigated. $\endgroup$ – Yves Daoust Nov 17 '20 at 13:53
2
$\begingroup$

Due to the fact that variables $x$ and $y$ appear as squared in the equation of $f$, one can consider that this curve is issued from this one

$$\begin{equation} g(x,y)\equiv -1296 x^2 \left(4 y+9\right)+9 x \left(-128 y^2+864 y+2187\right)-y \left(8 y+81\right)^2=0 \end{equation}$$

enter image description here

Curves with equations $f(x,y)=0$ (blue) and its mother curve $g(x,y)=0$ (red).

Let us call $g$ the "mother function" of $f$.

Important: we will consider the mother function/curve only in the first quadrant $x,y \ge 0$.

In fact, we are faced to a general generation scheme.

When we apply to the curve of $g$ the (inverse) transformation

$$(x,y) \to (X=\sqrt{x},Y=\sqrt{y}) \tag{1}$$

we obtain the upper right quarter of the lemniscate situated in the first quadrant. The remaining parts of the curve are obtained by symmetries wrt to the axes.

The important point is to take a mother curve $g$ in the first quadrant:

  • crossing the $x$ axis in two points: $x=0$ and $x=x_1$.

  • having a derivable cartesian equation $y=\gamma(x)$ with an increasing then decreasing part, more precisely such that there exists $x_0$ with $\gamma'(x)>0$ if $x \in [0,x_0)$ and $\gamma'(x)<0$ if $x \in (x_0,x_1]$.

In fact this is enough for having an image curve by (1) with a vertical axis and a maximum due to the fact that the jacobian matrix of transformation (1):

$$J=\begin{pmatrix}\dfrac{1}{2\sqrt{x}}&0\\0&\dfrac{1}{2\sqrt{y}}\end{pmatrix}$$

is such that $J_{22}$ tends to $\infty$ if $y \to 0_+$ and $J_{11}$ tends to $\infty$ if $x \to x_0$ (abscissa of the maximum of the mother curve).

Therefore, with these conditions, one can build many curves of type $g(x,y)=0$ yielding as many curves $f(x,y)=0$ having a lemniscate shape like this one.

Conclusion: For the reasons above, the given curve is a lemniscate among many others. I see no reason (in particular with its complicated coefficients) that it is amenable to one of the classical lemniscates (Bernoulli, Gerono, which are issued from second degree "mother" curves as shown on fig. 2 and 3).

enter image description here

The lemniscate of Gerono ($x^4-x^2+y^2=0$) is issued from a parabola ($x^2 -x +y=0$) by the process described in the text (transformation (1) applied to its first quadrant part, followed by symmetries).

enter image description here

The lemniscate of Bernoulli $(x^2+y^2)^2=x^2-y^2$ is also issued from a (45° slanted) parabola with equation $(x+y)^2=x-y$ by the same process.

Remark: Lemniscates are still studied. See for example this nice 2014 article.

$\endgroup$
1
  • $\begingroup$ I have added the case of the lemniscate of Bernoulli. Any comment ? $\endgroup$ – Jean Marie Nov 17 '20 at 13:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.